Echocardiographic Assessment of Cardiovascular
Hemodynamics in Normal Pregnancy
Dushyant K. Desai, MBChB, FCP, Jagidesa Moodley, FRCOG, and D. P. Naidoo, FRCP
OBJECTIVE: The factors affecting cardiac output in normal
pregnancy remain controversial. This study prospectively
evaluates maternal central hemodynamics and cardiac
structure and function by echocardiography, together with
maternal stature correction and correlation of these vari-
ables in healthy pregnant women in the latter half of
pregnancy.
METHODS: One hundred sixty echocardiographic studies
were performed in 35 healthy pregnant women for longi-
tudinal evaluation from early second trimester until term
and 6 –12 weeks postpartum.
RESULTS: Cardiac output increased significantly at the
early to mid third trimester and was maintained until
term. It increased predominantly in the latter half of preg-
nancy, and peak cardiac output of 46 –51% occurred from a
15% increase in heart rate and 24% increase in stroke
volume. Maternal cardiac output measured in the early
third trimester showed a good correlation with maternal
body surface area (r ⴝ 0.72; P < .001) and fetal birth weight
(r ⴝ 0.52; P ⴝ .008). Left ventricular systolic function was
preserved until term.
CONCLUSION: Maternal cardiac output peaks in the early
to mid third trimester and is maintained until term. Signif-
icant correlations were observed among maternal cardiac
output, maternal body surface area, and fetal birth weight.
(Obstet Gynecol 2004;104:20 –9. © 2004 by The Ameri-
can College of Obstetricians and Gynecologists.)
LEVEL OF EVIDENCE: II-2
Normal pregnancy is accompanied by maternal cardio-
vascular adaptations that include an increase in cardiac
output with a decline in blood pressure and systemic
vascular resistance. However, the precise changes of
cardiac output in normal pregnancy have remained con-
troversial. In a meta-analysis of cross-sectional studies by
van Oppen et al1 to evaluate a trend, the authors showed
large ranges in cardiac output across studies. Despite
numerous limitations of the above meta-analysis, pool-
ing data from each trimester showed a tendency to
From the MRC/UKZN Pregnancy Hypertension Research Unit and Department
of Obstetrics and Department of Gynaecology and Cardiology, Nelson R. Mandela
School of Medicine, University of KwaZulu-Natal, Durban, South Africa.
VOL. 104, NO. 1, JULY 2004
20 © 2004 by The American College of Obstetricians and Gynecologists.
Published by Lippincott Williams & Wilkins.
higher cardiac output in the second trimester compared
with the first and lower cardiac output in the third
trimester compared with the second. This observation is
supported by the Doppler echocardiographic study of
Hennessy et al2 that demonstrated a peak cardiac output
at 32 weeks of gestation of 49% that declined to a value of
21% at term.
Van Oppen et al1 also performed a meta-analysis of 6
longitudinal studies that had 2 or more cardiac output
measurements during pregnancy. The authors found
widely divergent changes in cardiac output between the
second and third trimesters, with 2 studies showing an
increase, 2 with no change, and 2 with a decrease. Both
van Oppen et al3 and Duvekot and Peeters4 cited patient
factors rather than technique as being responsible for the
apparent divergent trends of cardiac output in the third
trimester.
Although Thornburg et al5 in their review reported
that cardiac output peaks in the mid third trimester by
approximately 50%, the peaking of cardiac output has
been reported to occur at gestations varying from 24
weeks to term. In addition, the relative contributions to
cardiac output made by increases in heart rate and stroke
volume have not been well addressed. Furthermore,
most studies of maternal hemodynamics in normal preg-
nancy have reported measurement of cardiac output
rather than a stature-corrected measure of cardiac index.
This study evaluates echocardiographic maternal central
hemodynamics, cardiac structure and function, and ma-
ternal stature in healthy pregnant women.
MATERIALS AND METHODS
This study was conducted at the Obstetric Unit, King
Edward VIII Hospital, Durban, South Africa. Partici-
pants included healthy normotensive women with a
singleton pregnancy. They were selected and enrolled at
the antenatal clinic. Women with a history of any med-
ical disorder were excluded from the study. The Univer-
sity of Natal Ethics Committee granted ethical approval
for the study, and all participants gave informed consent.
0029-7844/04/$30.00
doi:10.1097/01.AOG.0000128170.15161.1d
 At enrollment, the attending obstetrician performed
an obstetric assessment together with clinical assessment
of gestational age. Obstetric ultrasound examination was
performed to obtain accurate fetal gestation, to exclude fetal
abnormalities, and to confirm a singleton pregnancy.
All participants had a clinical examination by 1 of the
authors (D.K.D.) to rule out preexisting heart disease.
Enrolled women rested in the left lateral position with
assistance of a foam wedge for 10 minutes. Echocardio-
graphic studies were performed thereafter (by D.K.D.)
using the Ultramark 9-HDI imaging system with a 2.5
MHz transducer (Scientific Medical Systems, Inc., New
York, NY). Two-dimensional echocardiography facilitated
accurate M-mode recordings and color flow mapping facil-
itated Doppler measurements according to standard crite-
ria.6 Ten normotensive participants at 32–36 weeks of
gestation had repeat echocardiographic studies 2 hours
apart to evaluate reproducibility; the mean percentage error
for cardiac output was measured at 2.0%.
In brief, the procedure was carried out as follows: an
initial 2-dimensional study in the standard parasternal
long axis and short axis planes followed by the apical
4-chamber plane view was performed to evaluate cardiac
structure and obtain a visual assessment of left ventricle
(LV) contractile function. Two-dimensional-imaging–
directed M-mode studies were then performed at the
level of aorta, left atrium, and left ventricle at mid
position between the tips of the mitral valve and papil-
lary muscle. Frozen M-mode images on screen were used
to measure chamber size and ventricular wall thickness.
Pulsed Doppler flow across the mitral valve was re-
corded just beyond the tips of the mitral valve leaflets to
obtain the LV diastolic filling pattern.
Heart rate and blood pressure (BP) were measured
during the echocardiographic study by an automated BP
measuring device (Critikon [SA], GE Medical Systems,
Melville, NY) at 3-minute intervals. Blood pressure was
measured in the dependent (left) arm. The average of
5–7 measurements obtained was accepted as representa-
tive heart rate and blood pressure measurements at
echocardiography. Mean blood pressure was computed
using the standard formula of mean BP (mm Hg) ⫽
diastolic BP ⫹ 1/3 (systolic BP ⫺ diastolic BP). These
values were used with echocardiographic-Doppler– de-
rived stroke volume to obtain cardiac output, and sys-
temic vascular resistance (SVR) was computed using the
following formula:
SVR (dyn ⫻ sec/cm5) ⫽ mean blood pressure in mm Hg/
cardiac output (I/min) ⫻ 80
Cardiac output derived from LV Doppler was ob-
tained by the standard accepted method as described by
VOL. 104, NO. 1, JULY 2004
Ihlen et al.7 Three estimates of the aortic annulus diam-
eter were made in the parasternal long axis plane, and
the average was accepted as the representative cross-
sectional area at the aortic annulus. The velocity time
integral Doppler measurement was performed in the
apical—5— chamber view at the level of the aortic valve
using pulsed Doppler. The best profile of 3 measure-
ments was selected and measured; the maximal velocity
time integral value was accepted. Calculation of LV
stroke volume was done using the following formula:
LV Doppler stroke volume (mL) ⫽ VTI aorta (m) ⫻
aortic annular area (mm2)
where VTI(m) ⫽ velocity time integral at the level of
aorta in meters.
The product of stroke volume and mean heart rate
(average of 5–7 measures) at echocardiography pro-
duced a measure of cardiac output. Cardiac output was
corrected for stature by the standard DuBois and DuBois
formula8 that used weight and height to derive body
surface area to obtain cardiac index (L · min⫺1 · m⫺2).
Systemic vascular resistance was computed from cardiac
output, and mean arterial blood pressure was obtained
from the average of 5–7 measurements at echocardiog-
raphy; systemic vascular resistance index was derived by
correction for body surface area.
The LV contractility (systolic function) was measured
from M-mode recordings of the LV in the short axis view
at a level just beyond the tips of the mitral valve leaflets.
On-screen measurements of the LV M-mode tracing
were used to measure LV cavity dimensions in systole
and diastole, from which indices of LV contractile func-
tion (fractional shortening, ejection fraction, velocity of
circumferential fiber shortening, and LV end systolic
wall stress) were computed. At least 2 M-mode tracings
of the LV were performed to ensure that cavity dimen-
sions were within 2 mm and that of ventricular wall
thickness to be within 1 mm of each other. A third trace
was performed if the first 2 M-mode traces did not
provide satisfactory measurements, and representative
values were obtained by consensus with a cardiology
colleague.
The LV ejection time, measured from the aortic ve-
locity time integral profile, was used to correct fractional
shortening to obtain the velocity of circumferential fiber
shortening, an index of LV contractile function.
LVID(d) ⫺ LVID(s)
LV-fractional shortening (%) ⫽ ⫻ 100%
LVID(d)
where LVID(d) is the left ventricle internal dimension in
diastole and LVID(s) is the left ventricle internal dimen-
sion in systole.
Desai et al Echocardiography in Pregnancy 21
LV-Ejection fraction (%) (Teicholz) ⫽
LVV(d) ⫺ LVV(s) ⫻ 100%
LVV共d兲
where LVV(d) is the left ventricle volume in diastole,
LVV(s) is the left ventricle volume in systole, and where
LV volumes are calculated by the Teichholz formula:
(7) ⫻ 共D兲 3
LV Volume ⫽
2.4 ⫹ D
where D is the LV minor axis dimension (in centimeters)
in diastole. The LV septum and posterior wall thickness
together with LV cavity size in diastole were used to
compute LV mass (in grams) according to the formula of
Devereux.6 The LV mass was corrected for maternal
size by dividing LV mass by body surface area (in square
meters) to obtain the LV mass index (in grams per
square meter). The LV mass was also corrected for
height to obtain LV mass/height (in grams per meter).
The LV mass was derived using measurements of LV
septum, posterior wall thickness, and LV internal diam-
eter. The formulae used to compute LV mass and LV
mass index are listed below.
LV mass (g) ⫽ 1.04[(LVID(d) ⫹ LV septum(d) ⫹
LV pw(d)]3 ⫺ [(LVID(d)3] ⫺ 13.6
where LVID(d) is the left ventricle internal dimension
(in centimeters) in diastole, LV septum(d) is left ventricle
septal thickness (in centimeters) in diastole, and LV pw
(d) is left ventricle posterior wall thickness (in centime-
ters) in diastole.
LV mass (g)
LV mass index (g/m2) ⫽
Body surface area (m2)
Left atrial size was assessed from the 2-dimensional-
view– directed M-mode measurements to obtain left
atrial and aortic root diameters from which a measure of
the left atrial to aortic size ratio was computed.
The LV diastolic filling velocities across the mitral
valve were obtained using pulsed Doppler in the apical
4-chamber view; recordings were made at a position just
distal to the mitral valve leaflets. The results were re-
corded as early filling velocity (in meters per second),
late filling velocity, and the early/late diastolic filling ratio
was computed.
All statistical comparisons were performed using the
SPSS 11.0 statistical package (SPSS Inc., Chicago, IL).
The primary comparative analyses of cardiovascular
hemodynamic and structural variables were made using
paired-sample t tests. Where appropriate, comparisons
of average values between periods were made either by
independent samples t tests or 1-way analysis of vari-
ance, and significance was set at P ⬍ .05. Cross tabula-
tion (␹2 test) was used for evaluation of noncontinuous
22 Desai et al Echocardiography in Pregnancy
Table 1. Patients Studied at Indicated Gestational Periods
(n ⫽ 35)
Number of patients
Gestational period studied (%)
Second trimester
Early (14–19 wk of gestation) 6 (17)
Mid (20–23 wk of gestation) 23 (66)
Late (24–27 wk of gestation) 29 (83)
Third trimester
Early (28–31 wk of gestation) 33 (94)
Mid (32–36 wk of gestation) 28 (80)
Late (37–40 wk of gestation) 14 (40)
Postpartum (6–12 wk) 27 (77)
variables. Bivariate correlation and linear regression
analyses were used to estimate relationships among mea-
sured cardiovascular variables.
RESULTS
The group consisted of 35 women, in whom a total of
160 echocardiographic studies were performed, 58 in the
second trimester, 75 in the third trimester, and 27 post-
partum as listed in Table 1. The baseline characteristics
expressed as mean ⫾ 1 standard deviation at entry into
the study were age 23 ⫾ 5 years; 63% (n ⫽ 22) were
nulliparous, and 37% (n ⫽ 13) were multiparous; gesta-
tion at entry 21 ⫾ 5 weeks; height 157 ⫾ 5 cm; weight
67 ⫾ 12 kg; and body surface area 1.68 ⫾ 0.15 m2.
Gestation at delivery was 38 ⫾ 2 weeks; birth weight was
2.9 ⫾ 0.6 kg, and 3 women (9%) had caesarian delivery.
Cardiac output in a subgroup of 10 women without
missing visits is compared with the full cohort in Table 3.
Table 2 details the hemodynamic changes in cardiac
output and systemic vascular resistance with corre-
sponding changes in heart rate, stroke volume, and mean
blood pressure. Cardiac output increased predominantly
in the latter half of pregnancy and continued to increase
and peak at term. However, statistically significant in-
creases (P ⬍ .05) were seen at the early third, late second,
and mid second trimester periods. Using postpartum
value as baseline, a 46% increase in cardiac output was
present. This maximal cardiac output occurred as a
result of a 15% increase in heart rate and a 24% increase
in stroke volume. This 24% increase in stroke volume
was derived at Doppler echocardiography by a 9% in-
crease in velocity time integral at the aortic annulus and
a 15% increase in aortic annulus cross-sectional area.
Heart rate showed a statistically nonsignificant de-
crease at the late third trimester compared with the mid
third trimester. Although stroke volume increased until
term, statistically significant increases were noted only at
the late third and late second trimester periods. A statis-
OBSTETRICS & GYNECOLOGY
Table 2. Longitudinal Changes in Normal Pregnancy
Systemic vascular
Heart rate Stroke volume Cardiac output Mean blood resistance
Gestation (wk) (bpm) (mL) (L/min) pressure (mm Hg) (dyn · s⫺1 · cm⫺5)
14–19 75 ⫾ 8 (10) 66 ⫾ 11 (⫺6) 4.96 ⫾ 0.5 (4) 74 ⫾ 9 (⫺5) 1,214 ⫾ 163 (⫺9)
20–23 76 ⫾ 9 (10) 74 ⫾ 11 (6) 5.60 ⫾ 0.8 (18) 71 ⫾ 6 (⫺9) 1,033 ⫾ 186 (⫺23)
P .054 .327 .295 .336 .497
24–27 80 ⫾ 8 (16) 75 ⫾ 11 (7) 5.94 ⫾ 0.9 (25) 70 ⫾ 8 (⫺10) 966 ⫾ 200 (⫺28)
P .010* .133 .009* .355 .024*
28–31 80 ⫾ 8 (16) 77 ⫾ 11 (10) 6.18 ⫾ 0.9 (30) 70 ⫾ 7 (⫺10) 923 ⫾ 140 (⫺31)
P .489 ⬍ .001* ⬍ .001* .317 .014*
32–36 81 ⫾ 8 (17) 80 ⫾ 13 (14) 6.42 ⫾ 0.9 (35) 72 ⫾ 6 (⫺8) 917 ⫾ 139 (⫺31)
P .383 .202 .156 .010* .219
37–term 79 ⫾ 7 (15) 87 ⫾ 17 (24) 6.94 ⫾ 1.8 (46) 74 ⫾ 7 (⫺5) 902 ⫾ 202 (⫺32)
P .098 .048* .267 .024* .250
Postpartum 69 ⫾ 11 (0) 70 ⫾ 12 (0) 4.75 ⫾ 0.7 (0) 78 ⫾ 9 (0) 1,336 ⫾ 218 (0)
P .007* ⬍ .001* ⬍ .001* .008* .001*
Data are presented as absolute mean ⫾ 1 standard deviation (percent change from baseline).
Percent change from baseline is calculated by comparing mean value for the measured parameter with the postpartum value.
P values indicate statistical change from preceding gestational period.
* Statistically significant.

tically significant maximal decrease in systemic vascular
resistance was observed at the early third trimester, and
mean blood pressure by comparison showed a statisti-
cally significant increase after the early third trimester.
Table 3 shows the stroke volume and cardiac output
in a subgroup of women (n ⫽ 10) who had echocardio-
graphic studies at all periods from the mid second trimes-
ter and were compared with the full group. The stroke
volume and cardiac output between these groups were
similar, without any statistically significant differences.
Paired t tests showed similar trends and P values at the
various periods in the subgroup and full group and thus
validated conclusions drawn from the full study group
members, who had missing visits. The P values for
cardiac output increases in the mid and late third trimes-
ter in the subgroup (P ⫽ .083 and P ⫽ .156, respectively)
indicate that cardiac output probably peaks at a period
between the early and mid third trimester and is main-
tained until term. A significant increase in stroke volume
at term was also noted for both the subgroup and full
group.
Table 4 details changes in cardiac output and cardiac
index together with weight and computed body surface
area. It is noted that both cardiac output and cardiac

Table 3. Comparative Longitudinal Changes of Full Study Group (n ⫽ 35) and Subgroup (n ⫽ 10) Without Missing Visits
in Latter Half Of Pregnancy
Stroke volume (mL) Cardiac output (L/min)
Gestation (wk) Full group Subgroup Full group Subgroup
14–19 66 ⫾ 11 (⫺6) ... 4.96 ⫾ 0.5 (4) ...
20–23 74 ⫾ 11 (6) 70 ⫾ 13 (⫺3) 5.60 ⫾ 0.8 (18) 5.64 ⫾ 0.8 (12)
P .327 ... .295 ...
24–27 75 ⫾ 11 (7) 69 ⫾ 8 (⫺4) 5.94 ⫾ 0.9 (25) 5.68 ⫾ 0.8 (23)
P .133 .500 .009* .276
28–31 77 ⫾ 11 (10) 78 ⫾ 11 (8) 6.18 ⫾ 0.9 (30) 6.25 ⫾ 0.8 (36)
P ⬍ .001* .007* ⬍ .001* .024*
32–36 80 ⫾ 13 (14) 80 ⫾ 14 (11) 6.42 ⫾ 0.9 (35) 6.61 ⫾ 1.1 (43)
P .202 .253 .156 .083
37–term 87 ⫾ 17 (24) 88 ⫾ 17 (22) 6.94 ⫾ 1.8 (46) 6.97 ⫾ 1.8 (51)
P .048* .003* .267 .156
Postpartum 70 ⫾ 12 (0) 72 ⫾ 18 (0) 4.75 ⫾ 0.7 (0) 4.61 ⫾ 0.9 (0)
P ⬍ .001* ⬍ .001* ⬍ .001* ⬍ .001*
Data are presented as absolute mean ⫾ 1 standard deviation (percent change from baseline).
Percent change from baseline is calculated by comparing mean value for the measured parameter to the postpartum value.
P values indicate statistical change from preceding gestational period by using paired t tests.
t test comparison of means of stroke volume and cardiac output for full group vs subgroup at each time period were not significant.
* Statistically significant.

VOL. 104, NO. 1, JULY 2004 Desai et al Echocardiography in Pregnancy 23

Table 4. Longitudinal Changes in Normal Pregnancy
Body surface Cardiac output Cardiac index
Gestation (wk) Weight (kg) area (m2) (L/min) (L · min⫺1 · m⫺2)
14–19 62 ⫾12 (2) 1.62 ⫾ 0.14 4.96 ⫾ 0.5 (4) 3.09 ⫾ 0.39 (4)
20–23 64 ⫾ 11 (5) 1.65 ⫾ 0.13 5.60 ⫾ 0.8 (18) 3.40 ⫾ 0.46 (15)
P .069 .054 .295 .395
24–27 65 ⫾ 11 (7) 1.65 ⫾ 0.14 5.94 ⫾ 0.9 (25) 3.63 ⫾ 0.50 (22)
P ⬍ .001* ⬍ .001* .009* .011*
28–31 67 ⫾ 12 (10) 1.68 ⫾ 0.15 6.18 ⫾ 0.9 (30) 3.67 ⫾ 0.35 (24)
P ⬍ .001* ⬍ .001* ⬍ .001* .023*
32–36 69 ⫾ 12 (13) 1.70 ⫾ 0.14 6.42 ⫾ 0.9 (35) 3.77 ⫾ 0.41 (27)
P .003* .004* .156 .247
37–term 72 ⫾ 14 (18) 1.72 ⫾ 0.17 6.94 ⫾ 1.8 (46) 4.00 ⫾ 0.76 (35)
P ⬍ .001* ⬍ .001* .267 .388
Postpartum 61 ⫾ 13 (0) 1.61 ⫾ 0.17 4.75 ⫾ 0.7 (0) 2.97 ⫾ 0.46 (0)
P ⬍ .001* ⬍ .001* ⬍ .001* ⬍ .001*
Data are presented as absolute mean ⫾ 1 standard deviation (percent change from baseline).
Percent change from baseline is calculated by comparing mean value for the measured parameter to the postpartum value.
P values indicate statistical change from preceding gestational period.
* Statistically significant.

index show similar statistical increases at the indicated
periods. Table 5 estimates relationships among cardiac
output, cardiac index, and maternal stature variables in
the early and mid third trimester and fetal birth weight.
Significant correlations were noted between cardiac out-
put and fetal birth weight and maternal stature variables
as indicated. Linear regression analysis showed that in
the early and mid third trimesters, weight best predicted
maternal cardiac output (r2 ⫽ 0.56 and 0.50, respectively).
Table 6 shows changes in cardiac structure and func-
tion variables of left atrial size together with left atrial to
aorta size ratio, LV early/atrial diastolic filling ratio, LV
mass, LV mass index, and LV systolic function reflected
by fractional shortening percentage. Both left atrial size
and left atrial/aorta size ratio showed a significant in-
crease at term, followed by a significant reduction in size
postpartum. The LV diastolic filling of early/atrial filling
ratio showed a nonsignificant decrease in the third tri-
mester.
Table 6 also shows significant increases in LV mass
and LV mass index that are maximal at term. The mean
LV mass index remained well below the arbitrary cutoff
level of 110 g/m2 to diagnose LV hypertrophy. The
study also showed good correlation of LV mass with
stature-corrected indices of LV mass index (r ⫽ 0.93),
LV mass/height (r ⫽ 0.99), and LV mass/height1.7 (r ⫽
0.99). A lower LV mass index at 14 –19 weeks of gesta-
tion compared with postpartum value (6 –12 weeks) is
noted and probably reflects that LV mass measured
6 –12 weeks postpartum had not returned to normal

Table 5. Correlation of Indices in Normal Pregnant Women

Cardiac output Cardiac index Fetal birth weight
Fetal birth weight
Early 3rd 0.52 (.008)* 0.16 (.438) ...
Mid 3rd 0.60 (.004)* 0.33 (.151) ...
Body surface area
Early 3rd 0.72 (.000)* 0.04 (.816) 0.60 (.002)*
Mid 3rd 0.64 (⬍ .001)* 0.13 (.523) 0.63 (.002)*
Weight
Early 3rd 0.73 (.000)* 0.08 (.654) 0.62 (.001)*
Mid 3rd 0.66 (⬍ .001)* 0.17 (.397) 0.60 (.004)*
Height
Early 3rd 0.34 (.061)* ⫺0.16 (.383) 0.45 (.023)*
Mid 3rd 0.24 (.233) ⫺0.08 (.694) 0.53 (.014)*
Body mass index
Early 3rd 0.69 (.000)* 0.15 (.424) 0.52 (.008)*
Mid 3rd 0.52 (.006)* 0.09 (.642) 0.35 (.124)
Early 3rd ⫽ early third trimester (28 –31 weeks of gestation); mid 3rd ⫽ mid third trimester (32–36 weeks of gestation).
Data are presented as r values (P values).
* Statistically significant.

24 Desai et al Echocardiography in Pregnancy OBSTETRICS & GYNECOLOGY

Table 6. Longitudinal Echocardiographic Structural and Function Changes in Pregnancy
Left ventricle
Left Left Left ventricle L ventricle Left ventricle systolic †
Gestation (wk) atrium atrium/aorta filling* mass (g) mass index (g/m2) function (%)
14–19 3.0 ⫾ 0.2 1.18 ⫾ 0.12 2.1 ⫾ 0.6 102 ⫾ 16 (⫺15) 63 ⫾ 10 (⫺15) 32 ⫾ 4
20–23 3.1 ⫾ 0.6 1.34 ⫾ 0.28 2.1 ⫾ 0.4 127 ⫾ 18 (6) 77 ⫾ 11 (4) 34 ⫾ 4
P .391 .428 .495 .038‡ .099 .014‡
24–27 3.2 ⫾ 0.5 1.38 ⫾ 0.23 2.1 ⫾ 0.5 124 ⫾ 23 (3) 75 ⫾ 12 (1) 34 ⫾ 4
P .088 .431 .303 .096 .089 .277
28–31 3.2 ⫾ 0.5 1.35 ⫾ 0.23 2.0 ⫾ 0.6 131 ⫾ 24 (9) 78 ⫾ 11 (5) 34 ⫾ 4
P .171 .335 .067 .210 .376 .032‡
32–36 3.2 ⫾ 0.5 1.37 ⫾ 0.23 1.8 ⫾ 0.3 139 ⫾ 30(16) 81 ⫾ 13 (9) 33 ⫾ 4
P .457 .244 .208 .052‡ .072 0.074
37–term 3.5 ⫾ 0.4 1.46 ⫾ 0.17 1.8 ⫾ 0.3 151 ⫾ 35(26) 87 ⫾ 15 (18) 35 ⫾ 4
P .016‡ .006‡ .309 .027‡ .048‡ .369
Postpartum 3.0 ⫾ 0.5 1.27 ⫾ 0.21 2.0 ⫾ 0.5 120 ⫾ 31 (0) 74 ⫾ 16 (0) 32 ⫾ 4
P .002‡ .005‡ .048‡ .001‡ .006‡ .130
Data are presented as absolute mean ⫾ 1 standard deviation (percent change from baseline).
Percent change from baseline is calculated by comparing mean value for the measured parameter to the postpartum value.
P values indicate statistical change from preceding gestational period.
* Reflected by early diastolic to atrial filling ratio.
†
Expressed as fractional shortening.
‡
Statistically significant.

prepregnant values; this limits an accurate assessment of
the extent of LV mass increase in our study.
DISCUSSION
Using 6 –12 weeks postpartum as baseline, our study
shows the expected increase in cardiac output in normal
pregnancy of 46 –51%. Approximately half of this in-
crease occurred by 28 weeks of gestation. Although we
show a significantly higher cardiac output in the third
trimester compared with the second, the observed in-
creases of cardiac output among early, mid, and late
third trimesters were not statistically significant. A large
variability in measured cardiac output at the late third
trimester, probably occurring from patient factors, did
not allow for confident conclusions to be made about the
precise changes in the late third trimester. However,
the data do show that cardiac output certainly peaks in
the early to mid third trimester and thereafter is either
maintained or possibly undergoes a minimal nonstatisti-
cal increase until term.
The literature has conflicting data on cardiac output
changes during pregnancy, particularly in the third tri-
mester.3,4,9 –11 Although the meta-analysis of cross-sec-
tional studies by van Oppen et al1 showed a trend to a
lower cardiac output in the third trimester compared
with the second, the authors observed large ranges in
cardiac output among the different studies that did not
allow for any firm conclusions. In evaluating 6 longitu-
dinal studies, van Oppen et al1 found that cardiac output
between the second and third trimesters plateaued, de-
creased, or increased. Of these, the 4 studies with com-
parable techniques still showed striking differences in the
course of cardiac output in the third trimester, with
Duvekot et al12 showing a decrease of 11.5%, no change
by Robson et al,13 and increases of 9.3% by Mabie et al14
and 16.4% by Thomsen et al.15
Although design differences and measurement tech-
niques among studies can explain some of the reported
differences in maternal hemodynamics in normal preg-
nancy, most researchers concur that patient factors
rather than measurement error are largely responsible
for discrepancies in reported studies. The longitudinal
study by Robson et al13 of 13 patients, cited as the only
true longitudinal study in the van Oppen meta-analysis,1
surprisingly showed a significant increase in cardiac out-
put of 75% of peak value by 12 weeks of gestation that
was followed by a very gradual rise to peak cardiac
output occurring at 24 –36 weeks and cardiac output
being maintained thereafter until term. It is to be noted
that after 16 weeks of gestation, the Robson study data
did not show any statistically significant increase in
cardiac output at any of the defined 4-week periods.
Our study shows a similar increase in mean cardiac
output until term as described by Mabie et al14 in their
longitudinal study of 18 normotensive women. How-
ever, although the Mabie study showed a peak cardiac
output at term, none of the increases in the comparative
periods (4 weeks) were significant. In addition, a similar
wide variation in measured cardiac output at the mid and
late third trimester was also noted. Our study, by con-
trast, shows a statistically significant increase in cardiac
output in the early third trimester over the late second

VOL. 104, NO. 1, JULY 2004 Desai et al Echocardiography in Pregnancy 25

trimester, data that clearly indicate that cardiac output at
the very least peaks in the early to mid third trimester.
The increase in cardiac output thereafter in the mid and
late third trimester in our study, although not statistically
significant, does indicate that at the very least, cardiac
output in the mid and late third trimester is maintained.
The ability of our study to show a statistically signifi-
cant increase in cardiac output at the early to mid third
trimester compared with the Robson and Mabie studies
described above probably reflects differences in ob-
served cardiac output increase in the first half of preg-
nancy. We did not directly measure cardiac output in
early pregnancy, but using postpartum values as base-
line, our data show that 40% of peak cardiac output
occurred at 24 weeks compared with the Robson and
Mabie studies described above, where more than 75% of
peak cardiac output occurred by 24 weeks. These differ-
ences of a “predominantly early” versus a “ predomi-
nantly late” increase in cardiac output in normal preg-
nancy are not easily explained. Although maternal body
surface area and cardiac output increase expressed as a
percentage in our study are similar to the study by Mabie
et al,14 the peak cardiac index of 4.8 ⫾ 0.8 L · min⫺1 ·
m⫺2 in the Mabie study is higher than 4.0 ⫾ 0.8 L ·
min⫺1 · m⫺2 in our study. In addition, the mean birth
weight in our study was 2.9 ⫾ 0.6 kg compared with
3.4 ⫾ 0.6 kg in the Mabie study. Despite comment in the
Mabie study of having cardiac output measurements
that are higher than most reported studies, the above
observations provide supportive evidence for a positive
association between maternal cardiac output and fetal
birth weight. In addition, they generate hypotheses, es-
pecially for comparative studies in different population
groups where apparent normal maternal health and
accepted normal fetal birth weights may not be so
and simultaneous uteroplacental evaluation may pro-
vide a more precise research definition of normal healthy
pregnancy.
Most of the heart rate increase in our study (60%)
probably occurred in the first trimester and that of stroke
volume in the third trimester, because changes in heart
rate in the second and third trimester were not signifi-
cant. It therefore appears that the mild increase in cardiac
output in early pregnancy is largely accounted for by an
increase in heart rate, and the progressive increase in
cardiac output thereafter in the latter half of pregnancy is
due to an increase in stroke volume. The apparent
bimodal peaking of stroke volume occurring initially in
the second trimester probably occurs as a result of
changes in heart rate described above. These observa-
tions are supported by the hypothesis of Carbillon et al16
that adaptation of vascular tone in early pregnancy pre-
26 Desai et al Echocardiography in Pregnancy
cedes and probably triggers blood volume and cardiac
output increase.
Noting a satisfactory association between maternal
cardiac output and maternal body size in our study,
some of the variations in cardiac output in normotensive
pregnant women can be explained by differences in
maternal stature and its surrogate marker of fetal birth
weight. Despite earlier reports of a poor correlation
between maternal cardiac output and body surface area
by van Oppen et al17, our study, together with reported
cardiovascular studies outside pregnancy in children and
obese adult participants by de Simone et al,18 indicate
that variables such as cardiac output need correction for
participants’ body stature for appropriate comparison
between groups. The need for correction of cardiovas-
cular variables for maternal size is highlighted in some
early studies. Rosso et al19 found that underweight
mothers had lower birth weight babies compared with
normal weight mothers, which corresponded with lower
maternal plasma volume and cardiac output. Similarly,
Carpenter et al20 found differences in maximal oxygen
uptake when evaluating effects of maternal weight on
exercise during pregnancy, differences that were not
present when oxygen uptake measurements were cor-
rected for maternal weight.
The poor correlation between cardiac output and
body surface area in the van Oppen et al17 study are not
easily explained; possible reasons include a bias from
including more women (n ⫽ 78) who had cardiac output
measured by thoracic electrical bioimpedance technique
than the 10 women from a cohort in a subfertility clinic
whose cardiac output was measured by the Doppler
technique (in whom 3 of the 5 antenatal evaluations were
in the first trimester). Both the Doppler and bioimped-
ance groups in the van Oppen study had cardiac output
increases of 88% and 92% of peak values by the fifth and
eighth weeks of gestation, respectively, observations that
do not make physiologic sense. Duvekot et al12 sug-
gested that this increase of cardiac output in the early and
mid first trimester is probably mediated by an endocrine
stimulus and hence, not surprisingly, poorly related to
maternal body surface area at this stage. It is also to be
noted that the correlation of maternal cardiac output
with maternal stature variables reported in our study
were made, albeit more appropriately, only in the third
trimester. Finally, despite criticism, the use of body sur-
face area computed by using the formula described by
DuBois and DuBois8 to standardize maternal cardiovas-
cular variables seems justified, because Wang et al21 in
their study evaluating predictors of body surface area
that included 60 women at gestation between 34 – 40
weeks concluded that several body surface area formu-
las, including the DuBois formulas, adequately predict
OBSTETRICS & GYNECOLOGY
measured body surface area over a wide range of patient
populations.
Differences in LV systolic function among studies can
explain some of the observed variation in measured
cardiac output and hence heart rate and stroke volume.
Normal pregnancy is associated with a significant drop in
blood pressure and systemic vascular resistance index, a
reflection of the reduced cardiovascular afterload and fall
in uterine vascular resistance. Physiologically, an in-
crease in preload and reduced afterload is often accom-
panied by an increase in left ventricular systolic function.
Our study demonstrates that LV fractional shortening
during pregnancy is preserved, findings that are similar
to data of Robson et al13 and Mabie et al.14 However,
these observations are at variance with studies by Mone
et al,10 who showed a transient fall in LV fractional
shortening during the third trimester, and that of
Schannwell et al,22 who also showed a reduction in left
ventricular systolic function values of fractional shorten-
ing and velocity of circumferential fiber shortening that
had a nadir at the first postpartum visit. These differ-
ences are not easily explained; the very plausible postu-
late both by Mone et al10 and Schannwell et al22 is that a
reduced preload in the latter part of the third trimester
contributes to a reversible fall in LV systolic function.
The study by Poppas et al23 is of particular interest, in
that they report a steady increase in cardiac output until
term but without any change in LV contractility, mea-
sured both by the load-dependent index of fractional
shortening and also by load-independent indices of LV
velocity of circumferential fiber shortening and LV sys-
tolic wall stress. As indicated, a limitation of LV frac-
tional shortening and echocardiography-derived ejection
fraction is that they are sensitive to loading conditions,
particularly preload and, to a lesser extent, afterload.
Noting this limitation, future studies evaluating maternal
LV contractility should focus more on the less load-
dependent indices such as LV velocity of circumferential
shortening and LV systolic wall stress. Group ethnic
profile as a factor associated with LV contractility is also
noted, because all participants in our study were Black
Africans. Desai et al24 had previously reported a very
high prevalence of peripartum cardiomyopathy in the
same population group. Such differences in LV systolic
function, although clinically small and transient, may
explain some of the divergent trends in cardiac output in
the third trimester.
In addressing a possible effect of parity on cardiac
output changes in pregnancy, Clapp et al25 in their
M-mode echocardiographic study suggested that cardio-
vascular adaptation is enhanced by a subsequent preg-
nancy. In the study by van Oppen et al3 that evaluated
cardiac output in the third trimester by thoracic electric
VOL. 104, NO. 1, JULY 2004
bioimpedance, the authors found a significant difference
in mean cardiac output between nulliparous and multip-
arous women. The number of participants in our study
did not allow for any confident conclusions on differ-
ences in hemodynamics between primiparous and mul-
tiparous normotensive pregnant women. Despite limita-
tions of measuring cardiac output in pregnancy by
M-mode echocardiographic and thoracic electrical bio-
impedance techniques, the above observations suggest
that some of the reported differences in reported cardiac
output among studies may be related to parity differ-
ences of study participants.
The variability of echocardiographic-Doppler– deter-
mined stroke volume and hence cardiac output depends
largely on measurement of aortic annulus diameter and
the computed aortic annulus cross-sectional area. Data
on appropriate and accurate aortic annulus measure-
ments are conflicting.5,7,14 In our study, using a single
operator, obtaining 3 measurements, and using the aver-
age as representative measurement, the aortic annular
measurement accounted for only 1.8% of the variation in
measured cardiac output. Robson et al26 have reported
within-subject, intraobserver, and temporal variation for
Doppler-derived cardiac output in pregnancy at less than
5%. Although a smaller number of subjects having echo-
cardiographic evaluation in the late third trimester did
not significantly affect comparative observations, it prob-
ably limited a more precise and confident measure of
changes between mid and latter portions of the third
trimester where expected absolute changes are smaller.
However, a limitation in our study is that reproducibility
was tested only at 32–26 weeks of gestation; this could
potentially be different at other gestational periods.
Our study shows a significant increase in LV mass and
LV mass index of 26% and 18%, respectively, using
6 –12 weeks postpartum as baseline, observations that
are similar to data of Mabie et al.14 However, reported
increases in LV mass in pregnancy have been variable,
being 10% in a study of 14 healthy pregnant women by
Poppas et al,23 16% in a study of 33 normotensive
women by Mone et al,10 and a 34% increase in left
ventricular mass index by Schannwell et al.21 These
variations in percentage increase in LV mass probably
reflect differences in the precise timing of postpartum
measurements, because Mone et al10 have indicated that
when evaluating LV structural changes, there is a lag
period of 1– 4 weeks to develop a compensatory increase
in LV mass to offset an increase in LV wall stress.
This study also shows that in healthy pregnant
women, there is an excellent correlation between LV
mass and its stature-corrected variables of LV mass
index, LV mass/height, or LV mass/height.1.7 An echo-
cardiographic diagnosis of LV hypertrophy (LV mass
Desai et al Echocardiography in Pregnancy 27
index ⱖ 110 g/m2) was noted in only 1 of 35 normoten-
sive women at term. The data of this study thus supports
the fact that although LV mass increases significantly in
normotensive pregnancy, levels that define the presence
of echocardiographic LV hypertrophy do not occur in
most women.
Our study shows significant increases in LV mass in
the mid and late third trimester that correspond to small
but statistically significant increases in mean blood pres-
sure, a significant increase in left atrial size at term, and a
nonsignificant decrease in LV diastolic filling value re-
flected by reduced LV early/atrial filling ratio. These
observations may be of relevance, because we27 have
reported abnormalities in LV diastolic filling in pregnant
women who presented with hypertensive crises compli-
cated by pulmonary edema. Because it is also accepted
that LV diastolic dysfunction often precedes systolic
dysfunction, differences in the degree of LV hypertro-
phy and diastolic filling could explain some of the re-
ported small differences in LV contractility and cardiac
output in the third trimester in normotensive pregnant
women.
In conclusion, our study shows a significant increase
in cardiac output at the early third trimester that is
maintained until term. Cardiac output increased pre-
dominantly in the latter half of pregnancy. We also show
a significant correlation between cardiac output and ma-
ternal stature and its surrogate marker of fetal birth
weight. Comparative analysis of our data with that of
Mabie et al14 further supports this association of mater-
nal cardiac output, stature, and fetal birth weight.
REFERENCES
1. van Oppen AC, Stigter RH, Bruinse HW. Cardiac output
in normal pregnancy: a critical review. Obstet Gynecol
1996;87:310 – 8.
2. Hennessy TG, MacDonald D, Hennessy MS, Maguire M,
Blake S, McCann HA, et al. Serial changes in cardiac
output during normal pregnancy: a Doppler ultrasound
study. Eur J Obstet Gynecol Reprod Biol 1996;70:117–22.
3. van Oppen AC, van der Tweel I, Alsbach GP, Heethaar
RM, Bruinse HW. A longitudinal study of maternal hemo-
dynamics during normal pregnancy. Obstet Gynecol
1996;88:40 – 6.
4. Duvekot JJ, Peeters LL. Maternal cardiovascular hemody-
namic adaptation to pregnancy. Obstet Gynecol Surv
1994;49 suppl:S1–14.
5. Thornburg KL, Jacobson SL, Giraud GD, Morton MJ.
Hemodynamic changes in pregnancy. Semin Perinatol
2000;24:11– 4.
6. Devereux RB, Pini R, Aurigemma GP, Roman MJ. Mea-
surement of left ventricular mass: methodology and exper-
tise. J Hypertens 1997;15:801–9.
28 Desai et al Echocardiography in Pregnancy
7. Ihlen H, Amlie JP, Dale J, Forfang K, Nitter-Hauge S,
Otterstad JE, et al. Determination of cardiac output by
Doppler echocardiography. Br Heart J 1984;51:54 – 60.
8. DuBois D, DuBois EF. A formula to estimate the approx-
imate surface area if height and weight be known. Arch
Intern Med 1916;17:863–71.
9. Varga I, Rigo J Jr, Somos P, Joo JG, Nagy B. Analysis of
maternal circulation and renal function in physiologic
pregnancies; parallel examinations of the changes in the
cardiac output and the glomerular filtration rate. J Matern
Fetal Med 2000;9:97–104.
10. Mone SM, Sanders SP, Colan SD. Control mechanisms
for physiological hypertrophy of pregnancy. Circulation
1996;94:667–72.
11. Ueland K, Metcalfe J. Circulatory changes in pregnancy.
Clin Obstet Gynecol. 1975;18:41–50.
12. Duvekot JJ, Cheriex EC, Pieters FA, Menheere PP,
Peeters LH. Early pregnancy changes in hemodynamics
and volume homeostasis are consecutive adjustments trig-
gered by a primary fall in systemic vascular tone. Am J
Obstet Gynecol 1993;169:1382–92.
13. Robson SC, Hunter S, Boys RJ, Dunlop W. Serial study of
factors influencing changes in cardiac output during
human pregnancy. Am J Physiol 1989;256 suppl:
H1060 –5.
14. Mabie WC, DiSessa TG, Crocker LG, Sibai BM, Arheart
KL. A longitudinal study of cardiac output in normal
human pregnancy. Am J Obstet Gynecol 1994;170:
849 –56.
15. Thomsen JK, Fogh-Andersen N, Jaszczak P, Giese J. Atrial
natriuretic peptide (ANP) decrease during normal preg-
nancy as related to hemodynamic changes and volume
regulation. Acta Obstet Gynecol Scand 1993;72:103–10.
16. Carbillon L, Uzan M, Uzan S. Pregnancy, vascular tone,
and maternal hemodynamics: a crucial adaptation. Obstet
Gynecol Surv 2000;55:574 – 81.
17. van Oppen AC, van der Tweel I, Duvekot JJ, Bruinse
HW. Use of cardiac index in pregnancy: is it justified?
Am J Obstet Gynecol 1995;173:923– 8.
18. de Simone G, Devereux RB, Daniels SR, Mureddu G,
Roman MJ, Kimball TR, et al. Stroke volume and cardiac
output in normotensive children and adults: assessment of
relations with body size and impact of overweight. Circu-
lation 1997;95:1837– 43.
19. Rosso P, Donoso E, Braun S, Espinoza R, Salas SP.
Hemodynamic changes in underweight pregnant women.
Obstet Gynecol 1992;79:908 –12.
20. Carpenter MW, Sady SP, Sady MA, Haydon B, Coustan
DR, Thompson PD. Effect of maternal weight gain during
pregnancy on exercise performance. J App Physiol 1990;
68:1173– 6.
21. Wang Y, Moss J, Thisted R. Predictors of body surface
area. J Clin Anesth 1992;4:4 –10.
22. Schannwell CM, Schoebel FC, Zimmermann T, Marx R,
OBSTETRICS & GYNECOLOGY
 Plehn G, Leschke M, et al. Left ventricular diastolic function
in normal pregnancy. A prospective study using M-mode
echocardiography and Doppler echocardiography [in Ger-
man]. Dtsche Med Wochenschr 2000;125:1069 –73.
23. Poppas A, Shroff SG, Korcarz CE, Hibbard JU, Berger DS,
Lindheimer MD, et al. Serial assessment of the cardiovas-
cular system in normal pregnancy: role of arterial compli-
ance and pulsatile arterial load. Circulation 1997;95:
2407–15.
24. Desai D, Moodley J, Naidoo D. Peripartum cardiomyop-
athy: experiences at King Edward VIII Hospital, Durban,
South Africa and a review of the literature. Trop Doct
1995;25:118 –23.
25. Clapp JF 3rd, Capeless E. Cardiovascular function before,
during, and after the first and subsequent pregnancies.
Am J Cardiol 1997;80:1469 –73.
26. Robson SC, Dunlop W, Moore M, Hunter S. Combined
Doppler and echocardiographic measurement of cardiac
output: theory and application in pregnancy. Br J Obstet
Gynaecol 1987;94:1014 –27.
27. Desai DK, Moodley J, Naidoo DP, Bhorat I. Cardiac
abnormalities in pulmonary oedema associated with
hypertensive crises in pregnancy. Br J Obstet Gynaecol
1996;103:523– 8.
Address reprint requests to: J. Moodley, Department of Ob-
stetrics and Gynaecology, Nelson R. Mandela School of Med-
icine, Private Bag 7, Congella, 4013, South Africa; e-mail:
gynae@nu.ac.za.
Received December 4, 2003. Received in revised form February 24,
2004. Accepted March 26, 2004.

THIRD EDITION OF WRITING GUIDE AVAILABLE

A Guide to Writing for Obstetrics & Gynecology was developed to help
prospective authors, especially those who are beginning in medical
journal writing, produce better papers.

Topics covered include:

• Planning and structuring a
manuscript
• Writing tips, including an
overview of syntax, grammar,
abbreviations, jargon and
acronyms, and troublesome
terms
• Common pitfalls of authorship
(eg, redundant publication and
conflict of interest)
• An overview of the peer-review
process

Copies of the third edition of this booklet are

available, free of charge, from the Editorial Office.
Requests may be submitted by FAX (202-479-0830)
or e-mail (obgyn@greenjournal.org).
OBSTETRICS&
GYNECOLOGY

VOL. 104, NO. 1, JULY 2004 Desai et al Echocardiography in Pregnancy 29